UX Collective: Dieter Rams' ten rules still fit designing for AI

Patrick Neeman’s June 2026 piece in UX Collective takes Dieter Rams’ ten principles — written in the 1970s for physical products at Braun, by a man who avoided computers throughout his career — and tests whether they still hold when applied to AI product design. The answer, Neeman argues, is yes, and in some cases they apply with more force than they did for hardware.

The central thread is restraint. Rams developed his principles as a counterweight to the tendency to add features rather than subtract them. Neeman’s argument is that this pressure is substantially more intense in AI product development, where capability is cheap and shipping fast creates incentives to accumulate rather than simplify.

The analysis works through each principle directly. Rams’ first rule — that good design should introduce genuine novelty — becomes a caution for AI products: AI should collapse multi-step workflows into fewer steps, not add visible capabilities that users must learn to use. “Good design makes a product useful” means testing whether AI saves real time in actual user workflows, not just in demos. “Good design is understandable” maps onto the obligation to signal AI limitations clearly — Neeman writes that a user who cannot tell where an answer came from also cannot tell when to trust it.

Two principles carry particular weight for AI product teams. “Good design is honest” addresses the mismatch between AI systems that write with confident tone and probabilistic outputs that carry genuine uncertainty. Confidence in prose is not the same as accuracy in fact, and designing this gap out is a meaningful design problem. “Good design is long-lasting” translates into building abstraction layers that allow the underlying AI model to be replaced without redesigning the interface — a practical architectural argument with direct product consequences.

“Good design is as little design as possible” — Rams’ shorthand for the less-but-better principle — Neeman applies to the feature accumulation common in AI products. The principle demands asking what to remove, not what to add.

The article is useful for product designers and team leads reviewing whether their AI features are genuinely serving users or accumulating for other reasons. Neeman does not prescribe specific frameworks or tools, staying at the level of principle — which means the piece ages well and applies across interface types, from chatbots to AI-assisted editing to generative design tools.